Example 7.1: In 2002, the NIH concluded that there was no benefit to taking estrogen after menopause.

Transcription

1 Chapter 7: Sampling

2 Example 7.1: In 1992, several major medical organizations concluded that women who took estrogen after menopause seemed to reduce their risk of heart attacks by 35 % - 50 %. In 2002, the NIH concluded that there was no benefit to taking estrogen after menopause.

3 Example: A typical hour of prime-time television shows three to five violent acts. Research shows that there is a clear association between time spent watching TV and aggressive behaviour caused by adolescents.

4 Example: A typical hour of prime-time television shows three to five violent acts. Research shows that there is a clear association between time spent watching TV and aggressive behaviour caused by adolescents. What are some lurking variables that may be confounded with how much TV he or she watches.

5 Your future is bright, very bright indeed, brighter than a bald man's dome. All images from Wikipedia.

6

7 The science of sampling Population: The entire group of individuals that we want information about. Sample: The part of the population that we actually examine in order to gather information.

8 Example: A mill produces large coils of thin steel for use in manufacturing home appliances. A quality engineer wants to submit a sample of 5cm squares to detailed laboratory inspection. She asks a technician to cut a sample of 10 such squares. Wanting to provide good pieces of steel, the technician carefully avoids the visible defects in the steel when cutting the sample. The lab results are wonderful, but the customers complain about the material they are receiving.

9 From on August 15, 2006.

10 The vast majority of economic experiments have had one particular social framing: the subjects are college students, playing against other college students, in a laboratory on campus. The ability to randomize subjects into treatment and control groups and to hold the experimental environment constant (or very nearly constant) has enabled experimenters to draw internally valid conclusions about the causal effects of different experimental procedures. But because experiments have tended to be limited to a particular subject pool in a particular social context, the extent to which their results generalize to other groups of people in other social contexts the external validity of the experiments remains open to question. Carpenter, Burkes, Verhoogen. Comparing Students to Workers: the effects of social framing on behavior in distribution games. Field Experiments in Economics (10) See also: List and Levitt. What do laboratory experiments tell us about the real world? Sep. 10, Henrich, et al. In search of Homo Economicus: Behavioral experiments in 15 small-scale societies. AEA Papers and Proceedings, May 2001.

11 Chemo found more harmful to breast cancer patients (AP) from on August 15, 2006.

12 A Simple Random Sample (SRS) of size n consists of n individuals from the population chosen in such a way that every set of n individuals has an equal chance of being the sample actually selected. Population: Matt Gayatri Cesar Neda Mustafa Possible Samples: Matt Matt Matt Matt Matt Matt Gayatri Gayatri Gayatri Cesar Gayatri Gayatri Gayatri Cesar Cesar Neda Cesar Cesar Neda Neda Cesar Neda Mustafa Neda Mustafa Mustafa Neda Mustafa Mustafa Mustafa

13 Choosing a SRS: 1. Use the Applet 2. Use a table.

14 Choosing a SRS: 1. Use the Applet 2. Use a table. TABLE B Random digits Line

15 There are 24 students in Statistics 1. Label individuals using as few digits as possible: 00, 01, 02, 03,..., 23

16 There are 24 students in Statistics 1. Label individuals using as few digits as possible: 00, 01, 02, 03,..., Enter Table B anywhere and read two-digit groups. To get a sample size of 12, read 12 valid two-digit groups.

17 TABLE B Random digits Line

18 A probability sample is a sample chosen by chance.

19 A probability sample is a sample chosen by chance. A stratified random sample is obtained by first dividing the population into groups of similar individuals (strata) and then choosing a SRS for each strata.

20 Example 7.7: Each state conducts an annual survey of seat belt use by drivers, following guidelines set by the federal government. The guidelines require probability samples. Seat belt use is observed at randomly chosen road locations at random times during daylight hours. The locations are not an SRS of all locations in the state but rather a stratified sample using the state s counties as strata.

21 Two sources of bias: undercoverage and nonresponse

22 Two sources of bias: undercoverage and nonresponse Current Population Survey: 6-7 % nonresponse

23 Two sources of bias: undercoverage and nonresponse Current Population Survey: 6-7 % nonresponse General Social Survey: 30% nonresponse

24 Two sources of bias: undercoverage and nonresponse Current Population Survey: 6-7 % nonresponse General Social Survey: Opinion Polls: 30% nonresponse 58% (?) nonresponse

25 Two more sources of bias: interviewer behaviour and question wording

26 Two more sources of bias: interviewer behaviour and question wording 4. In terms of HIV prevention work, does your organization: YES / NO Educate sex workers and partners about comprehensive HIV prevention? If yes, please describe. YES / NO Provide condoms to sex workers? If yes, how? If no, why? YES / NO Encourage sex workers to use condoms? If yes, how? YES / NO Promote empowerment of sex workers/trafficked persons? YES / NO Promote non-condom-related HIV prevention or family planning? YES / NO Other approaches to address the practice of safer sex. Please describe:

27 Two more sources of bias: interviewer behaviour and question wording 4. In terms of HIV prevention work, does your organization: YES / NO Educate sex workers and partners about comprehensive HIV prevention? If yes, please describe. From a survey by the YES / NO Provide condoms to sex workers? If yes, how? Equality If no, why? YES / NO Encourage sex workers to use condoms? If yes, how? YES / NO Promote empowerment of sex workers/trafficked persons? YES / NO Promote non-condom-related HIV prevention or family planning? YES / NO Other approaches to address the practice of safer sex. Please describe: Center for Health and Gender

28 Two more sources of bias: interviewer behaviour and question wording 4. In terms of HIV prevention work, does your organization: YES / NO Educate sex workers and partners about comprehensive HIV prevention? If yes, please describe. From a survey by the YES / NO Provide condoms to sex workers? If yes, how? Equality If no, why? The aim is to document the harmful YES / NO Encourage sex workers to use condoms? If yes, effects of these policies on the ground, how? and to conduct public education and advocacy around these effects. YES / NO Promote empowerment of sex workers/trafficked persons? YES / NO Promote non-condom-related HIV prevention or family planning? YES / NO Other approaches to address the practice of safer sex. Please describe: Center for Health and Gender

29 Question: How serious a problem do you think illegal immigration into the U.S. is? serious somewhat serious not very serious not serious at all

30 Question: How serious a problem do you think illegal immigration into the U.S. is? serious somewhat serious not very serious not serious at all From a survey by Negative Population Growth

31 Question: How serious a problem do you think illegal immigration into the U.S. is? serious somewhat serious not very serious not serious at all From a survey by Negative Population Growth Sample This report presents the findings of a nationally representative survey of 1,012 American adults (18+) who live in the continental United States. All sample was drawn randomly from an RDD (random digit dialing) list. Interviewing dates, method and timing All interviews were conducted by telephone from March 7-9, Each interview lasted approximately 7 minutes.

32 All interviews were conducted by telephone from March 7-9, Each interview lasted approximately 7 minutes. 36 % of respondents were from the South and about 20% from the other regions. Question: How serious a problem do you think illegal immigration into the U.S. is? serious somewhat serious not very serious not serious at all From a survey by Negative Population Growth Sample This report presents the findings of a nationally representative survey of 1,012 American adults (18+) who live in the continental United States. All sample was drawn randomly from an RDD (random digit dialing) list. Interviewing dates, method and timing

33 Senior Survey Report Westmont College Class of 2006 Since entering college have you been graded on a curve? 90 % Men 90 % Women answered frequently or occasionally

34 It is surprising to note that 90% of the seniors reported that they were graded on a curve. This result was unexpected based upon grade distribution reports available from the registrar indicating that the most frequent grades awarded to students are A s and B s. The traditional meaning of grading on a curve, where the most frequent grade assigned is a C, is clearly not understood by students as they responded to this survey item.